Sound mixing apparatus and method and multipoint conference server

A multipoint conference server connected to plural communication terminals via a network, the multipoint conference server comprising:

• a signal classifier that classifies, according to sampling frequencies, plural sound signals received from the plural communication terminals;
• plural mixing processing units that are provided according to sampling frequencies and mix sound signals having a same sampling frequency supplied from the signal classifier;
• and
• a speaker detector that selects at least one speaker sound signal including a sound utterance of a speaker from among plural sound signals inputted to the plural mixing processing units, wherein
• each processing unit of the plural mixing processing units adds up speaker sound signals selected by the speaker detector from among sound signals having the same sampling frequency inputted to the processing unit to generate a first added-up sound signal, converts the sampling frequency of the first added-up sound signal into a sampling frequency processable by the other mixing processing units, supplies the first added-up sound signals that are converted to the sampling frequency, to the other mixing processing units, and adds up, to generate a second added-up sound signal, the first added-up sound signal generated by itself and the first added-up sound signals that are converted to a sampling frequency processable by itself and that are supplied from the other mixing processing units wherein each of the plural mixing processing units generates, for each of the speaker sound signals selected by the speaker detector, a subtracted sound signal obtained by subtracting the speaker sound signal from the second added-up sound signal and encodes the subtracted sound signal.